INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

orgetown

-0.62

ape

-0.62

ter

-0.59

ucks

-0.59

edu

-0.58

ox

-0.58

âĶĢâĶĢ

-0.57

ipal

-0.57

jo

-0.57

adow

-0.57

POSITIVE LOGITS

issance

0.73

asion

0.71

 Suit

0.69

atchewan

0.66

essee

0.64

 Swap

0.63

 Wanted

0.63

PDATE

0.63

neys

0.63

anchester

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.